Commit Graph

11608 Commits

Author SHA1 Message Date
Maneesh Gupta d169ec0e8d Merge pull request #873 from gargrahul/fix_p2p_host_memory_access
Fixed issue of GPU device losing access to host pinned memory

[ROCm/clr commit: 2c610dde0b]
2019-01-22 10:53:18 +05:30
Maneesh Gupta d19169bcc8 Merge pull request #872 from yxsamliu/chrono
Use chrono instead of sys/time.h in test

[ROCm/clr commit: 325a5a2606]
2019-01-22 10:53:07 +05:30
Rahul Garg c6b09da295 Fixed issue of GPU device losing access to host pinned memory
[ROCm/clr commit: f8b399d3b5]
2019-01-22 06:25:43 +05:30
Yaxun Sam Liu 34f021d374 Fix get_time in tests for nvcc
[ROCm/clr commit: f49aed237b]
2019-01-21 14:07:50 -05:00
Yaxun Sam Liu 25b47bfed7 Use chrono instead of sys/time.h in test
sys/time.h is not available on Windows. Use C++11 <chrono> instead so that the tests
compile on Windows.


[ROCm/clr commit: 75a67c216f]
2019-01-20 00:05:27 -05:00
foreman b95b086b03 P4 to Git Change 1732914 by chui@ocl-promo-incrementor on 2019/01/19 03:00:08
SWDEV-2 - Change OpenCL version number from 2818 to 2819.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2566 edit


[ROCm/clr commit: dfbf0c56b3]
2019-01-19 03:11:47 -05:00
foreman 8cec62f4b5 P4 to Git Change 1732709 by gandryey@gera-w8 on 2019/01/18 14:18:03
SWDEV-79445 - OCL generic changes and code clean-up
	- Add  101010 GL interop formats mapping into CL_RGBA. The change will make sure the channel order consistency between OGL and OCL

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/opencl/amdocl/cl_gl.cpp#62 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldefs.hpp#47 edit


[ROCm/clr commit: 0b466ae6ff]
2019-01-18 14:41:37 -05:00
foreman d17f060d8f P4 to Git Change 1732709 by gandryey@gera-w8 on 2019/01/18 14:18:03
SWDEV-79445 - OCL generic changes and code clean-up
	- Add  101010 GL interop formats mapping into CL_RGBA. The change will make sure the channel order consistency between OGL and OCL

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/opencl/amdocl/cl_gl.cpp#62 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldefs.hpp#47 edit


[ROCm/clr commit: 02c5d38cb9]
2019-01-18 14:41:37 -05:00
foreman 2964973ee1 P4 to Git Change 1732420 by chui@ocl-promo-incrementor on 2019/01/18 03:00:08
SWDEV-2 - Change OpenCL version number from 2817 to 2818.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2565 edit


[ROCm/clr commit: cdbec540c4]
2019-01-18 03:41:54 -05:00
Maneesh Gupta 838baef822 Don't break cmake configure if roctracer headers not found
Change-Id: I35e49b52353ab24680836e397ab50136922f8641


[ROCm/clr commit: 3403527cfd]
2019-01-18 13:23:57 +05:30
foreman b38f1048aa P4 to Git Change 1732253 by asalmanp@asalmanp-ocl-stg on 2019/01/17 21:45:23
SWDEV-132899 - [OCL][GFX10] 70 subtests of Conformance Mipmaps (clCopyImage) test failed for image type 1Darray
	This is the follow up for CL#1517501

	copyImage1DA blit kernel uses image2d_array_t type for src/dst images. On gx10, num of arrays/layers is expected in Z component for a 2Darray image so a swap is required for 1Darray images when we use 2Darray image for the image copy. The copyImage1DA has code for swapping z and y components as follows:

	if (srcOrigin.w != 0) {
	coordsSrc.z = coordsSrc.y;
	coordsSrc.y = 0;
	}
	if (dstOrigin.w != 0) {
	coordsDst.z = coordsDst.y;
	coordsDst.y = 0;
	}

	So to use this path force the w component to 1 for src and dst images on gfx10 if image type is 1Darray.

	ReviewRequestURL = http://ocltc.amd.com/reviews/r/16538/diff/

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palblit.cpp#28 edit


[ROCm/clr commit: bd3af873ec]
2019-01-17 22:11:38 -05:00
foreman ece1f5716b P4 to Git Change 1732177 by gandryey@gera-w8 on 2019/01/17 18:08:26
SWDEV-174282 - [AMF] WIN10 Converter fails when scale YUY2 image with certain output width
	- When OCL creates an image view use the pitch value from the original surface

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#185 edit


[ROCm/clr commit: aab00a26bc]
2019-01-17 18:33:10 -05:00
foreman c7266c6fdc P4 to Git Change 1732168 by jatang@jatang_win_pal_lc on 2019/01/17 17:58:53
SWDEV-172202 - Workaround the scheduler for systems don't support PCIe 3 atomics properly.

	The idea is the scheduler uses a device side global as write_index, and only write the write_index back to the hsa queue when the last thread of the scheduler leaves.

	This change along with the library side change have been tested on systems with or without proper PCIe 3 atomics support.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocblit.cpp#29 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocsched.hpp#2 edit


[ROCm/clr commit: aa3989dcd0]
2019-01-17 18:20:24 -05:00
foreman ac188966d3 P4 to Git Change 1732008 by wchau@wchau_OCL_boltzmann on 2019/01/17 13:33:23
SWDEV-162389 - OpenCL Support for COMgr
	- added the machineTargetLC_ values, which was introduced in CL1702548, for Carrizo and Hawaii
	- requested by Joseph Greathouse for public users (https://github.com/RadeonOpenCompute/ROCm/issues/668)

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdefs.hpp#22 edit


[ROCm/clr commit: dc3e9a916d]
2019-01-17 15:20:06 -05:00
foreman 03fdc66db3 P4 to Git Change 1731717 by chui@ocl-promo-incrementor on 2019/01/17 03:00:10
SWDEV-2 - Change OpenCL version number from 2816 to 2817.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2564 edit


[ROCm/clr commit: 30a3da25f9]
2019-01-17 07:06:07 -05:00
Evgeny a4df59d819 fixing HSA_INIT_API cid args
[ROCm/clr commit: 47625cb8fd]
2019-01-16 23:45:44 -06:00
foreman 6f3d7bd508 P4 to Git Change 1731224 by jatang@jatang_win_pal_lc on 2019/01/16 10:40:44
SWDEV-172202 - Back out changelist 1730757.

	Failure in OCLDynamic tests in various TC Sanity tests.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.cpp#111 edit


[ROCm/clr commit: 1210448dd2]
2019-01-16 11:32:33 -05:00
Evgeny Mankov bc9b30f440 [HIPIFY][fix][#867] Relative paths routines fixes and improvements
+ Always check for existence any file or dir, first of all, because real_path and make_absolute differ: real_path additionally performs access on file/dir.
+ Relative path routines in source files implemented similarly to dirs.
+ Error handling improvements.


[ROCm/clr commit: 7bbf284636]
2019-01-16 19:30:09 +03:00
foreman 446d8c345c P4 to Git Change 1731071 by chui@ocl-promo-incrementor on 2019/01/16 03:00:08
SWDEV-2 - Change OpenCL version number from 2815 to 2816.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2563 edit


[ROCm/clr commit: a7d345bacc]
2019-01-16 04:07:29 -05:00
Maneesh Gupta 94af9b363c Merge pull request #864 from mangupta/swdev-157382
Switch to faster ocml variants for trig instrinsics

[ROCm/clr commit: 01f52762fc]
2019-01-16 10:17:20 +05:30
foreman 38a19685d9 P4 to Git Change 1730757 by jatang@jatang_win_pal_lc on 2019/01/15 15:41:42
SWDEV-172202 - Enable OpenCL 2.0 when PCIe Gen3 atomics is support.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.cpp#110 edit


[ROCm/clr commit: 8f937aef7c]
2019-01-15 16:36:10 -05:00
Maneesh Gupta e58fa4c026 Map more instrincis to ocml native
Change-Id: I54e55d5902931bafdabd0956e4e8c1d7b39f7173


[ROCm/clr commit: f74816efe6]
2019-01-15 19:40:37 +05:30
foreman d51893ff49 P4 to Git Change 1730491 by chui@ocl-promo-incrementor on 2019/01/15 03:00:07
SWDEV-2 - Change OpenCL version number from 2814 to 2815.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2562 edit


[ROCm/clr commit: 6b10f9322c]
2019-01-15 03:14:15 -05:00
Aaron Enye Shi 31e065ce9d Add backwards compatibility to dot funcs
[ROCm/clr commit: bdb84f3d9c]
2019-01-14 23:14:49 +00:00
Aaron Enye Shi 87ed1bcf69 Move fp16 dot functions to hip_fp16
[ROCm/clr commit: 7ba7d9a0f6]
2019-01-14 21:50:32 +00:00
Evgeny Mankov a13a08ba1e [HIPIFY] rename variables as requested in review
[ROCm/clr commit: e5e93a1088]
2019-01-14 21:19:34 +03:00
Evgeny Mankov e094dc49e9 [HIPIFY] Introduce '-D' option as hipify-clang's one
+ '-D' might be set as hipify-clang option (before separator '--' or without specifying separator at all);
+ '-D' as a clang option might be specified as well (after options separator '--');
+ All defines specified as hipify-clang defines (before options separator '--') will be set after clang's ones and will redefine the same clang's ones.


[ROCm/clr commit: 4b8fdda651]
2019-01-14 19:34:42 +03:00
Evgeny Mankov 99d41a38f0 [HIPIFY] Introduce '-I' option as hipify-clang's one
+ '-I' might be set as hipify-clang option (before separator '--' or without specifying separator at all);
+ '-I' as a clang option might be specified as well (after options separator '--').


[ROCm/clr commit: d5281a771a]
2019-01-14 18:23:54 +03:00
Maneesh Gupta f6d03c2c14 Switch to faster ocml variants for trig instrinsics
Change-Id: If62821e2fe1b0da91ad1b8c5580ebf1a009405e9


[ROCm/clr commit: 54e4463d16]
2019-01-14 13:01:57 +05:30
foreman 8ba1adb67c P4 to Git Change 1729594 by chui@ocl-promo-incrementor on 2019/01/12 03:00:05
SWDEV-2 - Change OpenCL version number from 2813 to 2814.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2561 edit


[ROCm/clr commit: ae5ae500b0]
2019-01-12 03:09:23 -05:00
foreman 6fba0b7aaf P4 to Git Change 1729577 by lmoriche@lmoriche_opencl_lc on 2019/01/11 22:21:53
SWDEV-145570 - [HIP] Windows build fix

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/build/Makefile.hip#9 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#10 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#11 edit


[ROCm/clr commit: 283c86ed26]
2019-01-11 22:45:33 -05:00
Aaron Enye Shi 896b320c5f Add Dot functions as amd_mixed_dot function
Introduce the Dot functions which are available in the device library. Forward those prototypes, and introduce HIP API to expose the usage of the dot functions.


[ROCm/clr commit: 84a028b9e1]
2019-01-11 23:31:02 +00:00
Evgeny Mankov aa02bb3ecb Merge pull request #861 from emankov/master
[HIPIFY][cmake] Add header files to the project explicitly

[ROCm/clr commit: a075eb8658]
2019-01-11 16:45:32 +03:00
foreman 1bf2990841 P4 to Git Change 1729095 by chui@ocl-promo-incrementor on 2019/01/11 03:00:06
SWDEV-2 - Change OpenCL version number from 2812 to 2813.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2560 edit


[ROCm/clr commit: a8fd09d6d9]
2019-01-11 03:20:28 -05:00
Maneesh Gupta c556d2f2cb Merge pull request #858 from lmoriche/Windows
Fix compilation issues with MSVC

[ROCm/clr commit: 7def56e23e]
2019-01-11 12:36:07 +05:30
Maneesh Gupta a60c85b353 Merge pull request #857 from ROCm-Developer-Tools/feature_ensure_lazy_init_initialises
Ensure that static structures are populated

[ROCm/clr commit: bc2f51a86f]
2019-01-11 12:35:52 +05:30
Maneesh Gupta 73fb88e87a Merge pull request #859 from LucasCampos/README-typo
Fix typo in README.md

[ROCm/clr commit: 6320a98498]
2019-01-11 07:54:19 +05:30
foreman 427070ef47 P4 to Git Change 1728676 by cpaquot@cpaquot-ocl-lc-lnx on 2019/01/10 14:29:52
SWDEV-145570 - [HIP] Add API tracing, enabled via LOG_LEVEL=3

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#19 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#42 edit
... //depot/stg/opencl/drivers/opencl/api/hip/trace_helper.h#1 add


[ROCm/clr commit: 427327b92c]
2019-01-10 14:46:01 -05:00
Evgeny Mankov ea1c394060 [HIPIFY][cmake] Add header files to the project explicitly
[ROCm/clr commit: b053cc0a21]
2019-01-10 19:45:43 +03:00
Evgeny Mankov c8d061ed14 [HIPIFY] Introduce '--' option - a separator between hipify-clang and clang options
+ Append '--' to the end of hipify-clang options in case of its absence.

[Reason]
+ Avoid the necessity for the user in JSON compilation database specification.
  Otherwise:
  Error while trying to load a compilation database:
  Could not auto-detect compilation database for file "test.cu"

  Though the hipification will be performed correctly without JSON database.


[ROCm/clr commit: 63b7fc5714]
2019-01-10 18:27:30 +03:00
Evgeny Mankov 175c294553 [HIPIFY] option cuda-path might be used with '--' as well as with a single dash: '-'
+ Update README.md


[ROCm/clr commit: 9b6af0a81d]
2019-01-10 18:01:33 +03:00
Lucas Campos ded3323a65 Fix typo in README.md
[ROCm/clr commit: ae4c2ee64c]
2019-01-10 09:45:23 +01:00
Laurent Morichetti c292915466 Fix compilation issues with MSVC
[ROCm/clr commit: 167bbffc4f]
2019-01-09 09:32:50 -08:00
Alex Voicu 47b142ad3b Ensure that static structures are populated
[ROCm/clr commit: 1c815170da]
2019-01-09 17:21:53 +00:00
Evgeny Mankov 7d81558657 [HIPIFY] Introduce CUDA installation path option '-cuda-path'
Repeats clang's '--cuda-path' option.

[Reason]
In case of absence of any other clang's options setting '-cuda-path' allows not to specify separator '--' before clang's '--cuda-path'.

+ Tests and scripts are updated accordingly.


[ROCm/clr commit: c4268855e3]
2019-01-09 20:18:36 +03:00
Evgeny Mankov d424128b31 [HIPIFY][cmake][win] Add cmake versioning depending on Visual Studio version
[ROCm/clr commit: 558894ab06]
2019-01-09 16:44:12 +03:00
Evgeny Mankov 587969c10e Merge branch 'master' of https://github.com/ROCm-Developer-Tools/HIP
[ROCm/clr commit: 5dba4e74cb]
2019-01-09 14:03:40 +03:00
Evgeny Mankov eb42f06a52 [HIPIFY][cmake] Add cmake versioning depending on CUDA version
[ROCm/clr commit: d160c7285c]
2019-01-09 13:59:48 +03:00
foreman eaa1f3b1e8 P4 to Git Change 1727825 by chui@ocl-promo-incrementor on 2019/01/09 03:00:06
SWDEV-2 - Change OpenCL version number from 2811 to 2812.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/utils/versions.hpp#2559 edit


[ROCm/clr commit: ca14911693]
2019-01-09 03:12:22 -05:00
Wen-Heng (Jack) Chung 2f64c76789 Introduce hash key to HIP module implementation
A hash calculated via FNV-1a algorithm is introduced in ihipModule_t, the
internal of hipModule_t. The hash is used by HIP module APIs such as

- read_agent_global_from_module

to determine whether the agent-scope globals for a module have been iterated.

This commit fixes one issue that applications which load / unload modules
frequently would occasionally fail. After deep investigation of the issue it
turns out the old implementation in read_agent_global_from_module uses
hipModule_t as the key, which is not robust enough, as hipModule_t instances
are allocated dynamically so there are cases that one memory address may be
used by multiple hipModule_t instances. The real solution is to introduce a
uniquely identifiable hash for the code object associated with the HIP module.
And that's the rationale behind this commit.


[ROCm/clr commit: 6e68d44220]
2019-01-08 17:33:40 +00:00